We propose to continue our studies of the structure and assembly of cytoskeletal filaments, in three areas. (1) The mechanism of nucleation of microtubule assembly will be investigated by fitting experimental assembly kinetics to curves generated by computer for different nucleation models. We have recently completd the fitting for one buffer system using a model comprising three phases: nucleation, lateral growth, and elongation. We now propose to study other buffer conditions, non-hydrolyzable analogues of GTP, and the effect of MAPs to determine how the nucleation pathway is affected by different assembly conditions. Electron microscopy will be used to follow all assembly reactions, to confirm that the final polymers and small intermediates are consistent with the computer model. (2) We will use electron microscopy to follow the pathway of assembly of keratin and desmin. Preparations of soluble subunits will be characterized and specimens will be prepared at short time intervals after initiating assembly. If we can visualize some of the small intermediates we should obtain important new information about the pathway of assembly and also about the structure of the complete intermediate filament, which is still largely unknown. (3) We will use small angle x-ray scattering to demonstrate the existence of and to characterize weak association complexes (lengthy collisions) of protein molecules. These complexes have been proposed to explain the very rapid kinetics of protein-protein interaction, but have not yet been demonstrated by physical measurement. They are thought to play a fundamental role in all protein association and assembly reactions, and may influence the structure of the cytoplasm. Our x-ray measurements should demonstrate the complexes if they exist, and should allow us to investigate the nature of the weak association with a variety of proteins and different solution conditions.